the Thermo-Electric Properties of Liquids. 541 



rent by the mere contact alone, and when this contact effect of a 

 metal with a liqnid differs in strength with two different metals, such 

 metals, each at the same temperature, and immersed in a liquid of that 

 temperature (whether cold or hot, and without the aid of chemical 

 action), generates a current, provided the contact effect of the metals 

 themselves does not neutralise that of the metals and liquid. 



Such an effect has been already obtained. Becquerel, for instance, 

 immersed pure platinum and gold in various acids, ammonia, and 

 neutral saline solution, and obtained very feeble currents. He con- 

 cluded that the state of the surface of the metals had great influence, 

 and that consequently capillary affinity played a very important part 

 in producing the effects. (See " Comptes Rendus Acad. Sci," Paris, 

 1870, pp. 480, 961, 1313; also "Chemical News," vol. xxii, p. 21, 

 and vol. xxiii, p. 222.) As neither gold nor platinum is perceptibly 

 corroded by dilute acids, nor usually by saline liquids, at ordinary 

 atmospheric temperatures, any electric current produced by such a 

 combination can only be due to contact, either of the metals with each 

 other or with the liquid. 



According to these facts, as in the thermo-electric apparatus there 

 is no contact of dissimilar metals, both the contact of the cold metal 

 (as well as that of the hot one) with the liquid, tends to generate a 

 current, and rise of temperature makes the warmer metal only more or 

 less electro-positive than it was previously. Another reason also for 

 concluding that both the cold and the hot piece of metal, at each 

 different temperature, tends to generate a current, is, that with two 

 pieces of the same metal at 60° F. in the apparatus, if one was made 

 colder, a current was produced in the same direction as when the other 

 was heated. The electrical effect, therefore, is a result of a difference 

 of action at the surfaces of contact of the two pieces of metal with the 

 liquid ; and the apparatus and method may be employed for detecting 

 differences of electric potential between metals and liquids at different 

 temperatures. 



In addition to Volta and Becquerel, others, viz., Pfaff, Fechner, Peclet, 

 Matteucci, Kohlrausch, Marianini, Hankel, Zamboni, Buff, and Thom- 

 son, have investigated the contact theory of voltaic action, and Clifton 

 has examined the difference of electric potential produced by the 

 contact of metals and liquids (see " Proc. Roy. Soc," 1877, vol. 

 xxvi, p. 299), but the metals he employed were easily corrodible ones, 

 viz., zinc, iron, and copper. 



As in my experiments the origin of the current appears to lie in a 

 conversion of heat into electricity, and heat probably disappears in the 

 process, it is a reasonable inference that, in the above experiments of 

 Becquerel's, &c, heat also disappears and produces the currents; and, 

 if this be correct, then it is also likely that, when a piece of metal is 

 simply immersed in a suitable liquid, the temperature is altered ; and 



